Over the last four years, 180 youths discharged from an impatient psychiatric facility have been successfully recruited and followed longitudinally as part of a FIRST Award. The project was designed to examine the course, predictors, and covariates over time of suicidal behavior among a high risk population of adolescents. Eighty four percent of the eligible sample of consecutively referred adolescents were recruited, and the cumulative rate of attrition in the study has been limited to less than 10%. This continuation application describes a proposal to follow these high risk youths for an additional five years until the youngest subjects reach late adolescence and the oldest subjects reach young adulthood. Late adolescence is a period of multiple developmental transition, and a period of time during which young people presumably are at continued risk for suicidal behavior. Within this developmental context are a number of psychiatric, cognitive, sociodemographic, and environmental factors that may interact in increasing the risk that youths will make additional suicide attempts. The first specific aim of this continuation is to examine the rates of suicide attempts among these high-risk adolescents from mid-adolescence to young adulthood. The second specific aim is to examine the effects of the course of affective disorder, and of comorbid affective and substance disorders on the likelihood of suicidal behavior. The final aim of this continuation is to test the utility of a hopelessness diathesis stress model (derived from Beck's cognitive model of depression) in predicting suicide attempts among high-risk youths.